Statistical aspects of experimental studies with complex mixtures

Some formal concepts are introduced relating to the statistical design and analysis of experiments with complex mixtures. Aspects relating to the chemical analysis of complex mixtures, identification of their major components, continuous monitoring of samples or related issues are not addressed. As a surrogate for experimental studies in general, the discussion is oriented towards the situation of long-term animal experiments. When complex mixtures of unknown components are to be tested, samples collected under a standardized protocol must be applied to the different exposure groups in such a way that the experimental groups differ only in the dose of the exposure they receive. If the composition of a mixture is known, selected components may be combined in experimental testing in order to study the nature of their joint effect. Careful consideration guided by a priori knowledge about the individual compounds' dose-response relationships is required to design and analyse such experiments. There is an apparent need to avoid confusion by unifying the terminology in this field.